CN106888504B - Indoor position fingerprint positioning method based on FM and DTMB signals - Google Patents

Abstract

The invention discloses an indoor position fingerprint positioning method based on a frequency modulation signal (FM) and a terrestrial digital television broadcasting signal (DTMB). The method comprises the following steps: selecting sampling points according to indoor environment, selecting different frequencies of FM and DTMB signals as reference frequencies, recording the intensity information of signals on the reference frequencies of the sampling points, and constructing a received signal intensity indication (RSSI) fingerprint database; and acquiring RSSI fingerprint information of the to-be-positioned site FM and DTMB signals, and matching the optimal position information according to a joint positioning algorithm. The invention uses FM and DTMB signals with stable signal strength as positioning signals, and has wide coverage range and low maintenance cost; the invention adopts different frequencies as references, overcomes the influence of the movement of personnel and objects on the positioning precision, and improves the anti-interference performance of the system; the combined positioning algorithm provided by the invention increases the diversity of positioning information, effectively counteracts partial errors and realizes indoor high-precision positioning.

Description

Indoor Location Fingerprint Positioning Method Based on FM and DTMB Signals

technical field

The invention relates to a high-precision positioning method for determining the indoor environment space position by using position fingerprints based on FM and DTMB signals.

Background technique

At present, with the rapid development of wireless communication technology and the popularization of smart mobile terminals, location-based services (LBS) are also developing continuously. In outdoor situations, the satellite positioning system (GNSS) has become the mainstream positioning technology due to its advantages of high precision (generally, the typical error does not exceed 10m) and low cost. In terms of indoor positioning, due to building occlusion, multipath interference, and low GNSS signal strength, satellite positioning technology based on time of arrival or angle of arrival will produce large multipath errors in indoor positioning. In an indoor environment, each point in the space has a unique signal strength corresponding to it. The position fingerprint positioning technology based on received signal strength (RSSI) utilizes the specificity of indoor environment to signal fading, and can achieve accurate indoor positioning by comparing the measured signal strength with the prior signal strength in the database. Considering factors such as signal coverage areas and deployment costs, Wi-Fi signals have become the preferred signal for location fingerprint positioning technology.

However, the research found that Wi-Fi signals also have great deficiencies in positioning practice. First, the coverage of a single Wi-Fi signal source is limited, and location fingerprint positioning technology requires a certain number of signal sources (not less than three) to ensure positioning accuracy, which greatly increases the cost of device deployment. Secondly, the frequency of the Wi-Fi signal is around 2.4GHz, which coincides with the resonance frequency of water, and 70% of the human body is water, which makes the strength of the received Wi-Fi signal easily affected by the movement of people and even the weather , thereby interfering with the positioning results. Therefore, many foreign researchers are currently trying to use FM radio instead of Wi-Fi as a positioning signal. The FM signal strength is stable, the coverage is very wide, and its frequency range is not easily disturbed, so it is a very ideal positioning signal.

Relevant foreign researchers have been able to use FM signals alone to achieve positioning accuracy of the same magnitude (<10m) as Wi-Fi signals. At present, there are few domestic reports on the use of FM signal positioning, which is mainly affected by different national conditions at home and abroad. There are many and widely distributed FM signal sources in the same area abroad, which helps to improve the accuracy of FM indoor positioning. In my country, FM broadcasting is uniformly transmitted by designated transmitting stations, and the FM signals that can be received in the same area are all transmitted by the same signal transmitting tower, which makes the accuracy of indoor positioning using FM signals in my country low.

In view of the deficiency of Wi-Fi indoor positioning and the national conditions of FM signal transmission in my country, the present invention proposes to simultaneously use digital television terrestrial broadcasting (DTMB) signals and FM signals to form a new joint positioning system to improve the accuracy of indoor positioning. The DTMB signal is sent by the TV signal tower, and its frequency is much higher than that of the FM signal. Taking Tianjin as an example, the center frequencies of the two frequency bands of the DTMB signal are 660MHz and 740MHz, and the wavelength is between 0.41m and 0.45m. Electromagnetic wave signals of this wavelength will be refracted at the corners of buildings. Therefore, when the positioning target turns around the corner, the propagation length of the DTMB signal will change significantly, and the RSS value of the signal will also change greatly at this time, which is conducive to positioning the target. Considering the complex situation in the building, the introduction of DTMB signal will greatly improve the accuracy of FM signal positioning.

Contents of the invention

In view of the defects of the above-mentioned various methods, the present invention proposes an indoor position fingerprint positioning method based on FM and DTMB signals, which overcomes the limitations and instability of traditionally using Wi-Fi signals for positioning, and solves the problem of using FM signals as The problem of low accuracy of a single positioning signal can achieve indoor high-precision positioning.

Realize the technical scheme of the present invention as follows:

The method determines the spatial position of the point to be located in the indoor environment by receiving the strength information of different frequencies of FM and DTMB signals. The specific steps are:

(1) Select sampling points according to the indoor environment, and select different frequencies of FM and DTMB signals as reference frequencies;

(2) Utilize the antenna to receive FM and DTMB signals at each sampling point, record the strength information of the signal on the reference frequency, and construct the RSSI fingerprint database of FM and DTMB signals;

(3) Utilize the antenna to receive FM and DTMB signals at the point to be located, record the strength information of the signal on the reference frequency, and obtain the RSSI fingerprint information of the FM and DTMB signals at the point to be located;

(4) Through the joint positioning algorithm, the RSSI fingerprint information of the FM and DTMB signals of the point to be located is matched with the RSSI fingerprint database to obtain the positioning result.

Further, the FM signals described in the present invention are uniformly transmitted by designated transmitting stations, and the receivable FM signals in the same area all come from the same signal transmitting tower, and the FM signal frequency ranges from 87.5MHz to 108MHz.

Further, the different frequencies of FM and DTMB signals in the step (1) of the present invention are selected according to the channel environment. Both FM and DTMB signals contain multiple frequencies, and the signal strength on each frequency is not suitable for positioning in a specific channel environment, and if the strength information of all frequencies is put into the RSSI fingerprint database, the amount of calculation will be increased. It will also affect the matching of fingerprint information in the later stage and reduce the efficiency of the system. So the present invention converts each frequency value of FM and DTMB signal into wavelength according to the difference of penetration loss and diffraction loss of obstacles to obstacles of different wavelengths, and the conversion formula is: λ=c/f, where c is the speed of light, Its value is 299792458m/s. Select the signal whose wavelength is larger than the maximum size of the obstacle and the signal whose wavelength is smaller than the smallest size of the obstacle for optimal combination.

Further, the strength information of the different frequencies of FM and DTMB signals described in the present invention is represented as a vector, and N sampling points are selected in the area, and the signal strengths of the P reference frequencies that are measured at each point are received, then the rth The signal strength of the sampling point is expressed as:

为第i个频率进行Q次测量后的平均值，即/>

in

The average value of Q measurements for the i-th frequency, i.e. />

Further, the joint positioning algorithm in step (4) of the present invention combines two kinds of mutually independent information, deterministic information and probabilistic information. The deterministic information of the positioning result is obtained from the neighbor reference point; the probabilistic information of the positioning result is obtained from the probability distribution of the signal strength. The two kinds of information are assigned different weight values to obtain the final positioning result. The specific formula is:

loc＝ω ₁ loc ₁ +ω ₂ loc ₂

Among them, loc ₁ is the deterministic information of the positioning result, loc ₂ is the probabilistic information of the positioning result, and ω ₁ and ω ₂ are the weights of the two kinds of information respectively.

Further, the deterministic information described in the present invention is obtained by comparing the real measurement value with the prior measurement strength of the position, and its specific implementation method is: evaluate the Euclidean distance from the point to be located to all sampling points, and select The nearest K distances, the points corresponding to these K distances are the neighbor reference points. At the same time, considering the difference between the K distances and the real-time measured signal strength, different weight values are assigned to the K neighboring reference points.

Further, the probabilistic information described in the present invention is estimated based on a probability model, and its specific implementation method is: when measuring the FM and DTMB signal strength of the sampling point, the signal strength of the point is measured multiple times and the signal strength of the point is measured. The distribution of signal strength is estimated; when positioning, the position of the point to be positioned is regarded as an independent random variable, and the probability of occurrence of the measured signal strength is analyzed to estimate the position with the highest probability of occurrence of the measured strength as the positioning result.

According to the Bayesian formula, when the measured signal strength is rss, the probability P(LOC _r |rss) of the positioning position at LOC _r is:

Among them, P(rss) is the probability that the measured value is rss, P(LOC _r ) is the probability that the point to be located appears in the LOC, and P(rss|LOC _r ) is the probability that the rss intensity appears at the LOC _r point.

Use the Gaussian distribution to fit the distribution of the signal strength of each sampling point:

Among them, Q is the number of samples actually measured, and RSS _rq obeys the normal distribution of N(rss _r , σ).

Compared with the existing technology, the indoor positioning method proposed by the present invention has the following advantages:

(1) The positioning signals FM and DTMB signals used in the present invention have wide coverage, stable signal strength, and less susceptible to interference in the frequency range. Compared with Wi-Fi signals, the positioning stability of the method of the present invention is stronger and the maintenance cost is lower. Low; compared with using FM signals alone, the positioning accuracy of the method of the present invention is higher.

(2) When constructing the RSSI fingerprint database of FM and DTMB signals, through the selection of different frequencies of the signals, while simplifying the workload of the system, it overcomes the impact of the movement of people and objects on the positioning accuracy of the system, and improves the anti-interference of the positioning method performance.

(3) The present invention combines deterministic information with probabilistic information. The two information mechanisms are different, and the errors of the two are also independent of each other. The joint use of these two types of information can effectively increase the diversity of positioning information and effectively offset part of the error. Improve the stability of positioning method.

(4) The method adopted in the present invention is applicable to any environment where FM and DTMB signals can be received simultaneously, and overcomes the disadvantages that the existing method is only limited to indoor environments and has a small coverage area.

Description of drawings

Fig. 1 is the flow chart of the present invention's indoor position fingerprint location method based on FM and DTMB signal;

Fig. 2 is a difference diagram of signal propagation paths of different wavelengths;

Fig. 3 is a flowchart of the joint positioning algorithm.

Detailed ways

The method of the present invention will be described in detail with reference to the drawings and embodiments.

As shown in Figure 1, the indoor location fingerprint positioning method based on FM and DTMB signals, the specific process is as follows:

(1) Select sampling points according to the indoor environment, and select different frequencies of FM and DTMB signals as reference frequencies.

For a certain indoor positioning environment, considering the specificity of the signal strength of each point in the space and the actual positioning requirements, a two-dimensional modeling of the space is carried out, and then the environment is divided equidistantly to form a grid form. Depends on the structure and specific positioning requirements. Take the center position of the grid as the sampling point, establish a coordinate system, obtain the two-dimensional position coordinates (x, y) of the sampling point, and correspond the serial number of the sampling point to the position one by one, then the rth sampling point is expressed as: P _r (x _r , y _r ), r=1, 2, ..., N, N is the total number of sampling points.

The selection of FM and DTMB signal reference frequencies is an important prerequisite for constructing a reasonable and efficient RSSI fingerprint database. Both FM and DTMB signals contain multiple frequencies, and not the signal strength on each frequency is suitable for a specific channel environment, and if the strength information of all frequencies is put into the RSSI fingerprint database, the amount of calculation will be increased, and the It will affect the matching of fingerprint information in the later stage and reduce the efficiency of the system. Therefore, the present invention selects a frequency with a large difference in the receiving range from the frequencies of FM and DTMB signals as the reference frequency according to the differences in the penetration loss and diffraction loss of obstacles with different wavelengths.

The frequency values of FM and DTMB signals are converted into wavelengths, and the conversion formula is: λ=c/f, where c is the speed of light, and its value is 299792458m/s.

In an indoor environment, the size of obstacles is usually 2 to 3 meters. As shown in Figure 2, when signals of different wavelengths encounter obstacles, their propagation paths are different. In the present invention, the frequency range of the FM signal is from 87.5MHz to 108MHz, then the wavelength range is between 2.8m to 3.5m, and the electromagnetic wave of this length will not produce refraction when encountering an obstacle less than 3 meters, and the signal The loss is small, and the signal strength change is not obvious enough, which is not conducive to accurate positioning. The frequency of DTMB signals is much higher than that of FM signals. Taking Tianjin as an example, the center frequencies of the two frequency bands of DTMB signals are 660MHz and 740MHz, and their wavelengths are between 0.41m and 0.45m. Refraction will occur when there is an obstacle of 2 meters, the signal loss will be greater, and the signal strength will vary greatly.

Because the indoor channel environment is very complex, when selecting signals of different wavelengths, too many or too few will have a bad influence on the positioning results. After many experiments, the positioning effect is the best when the combination of the signal with the wavelength larger than the maximum size of the obstacle and the signal with the wavelength smaller than the minimum size of the obstacle is selected. Therefore, the present invention selects a part of reference frequencies in FM and DTMB signals respectively, and then combines the two. The selected reference frequencies are expressed as F _l , i=1, 2, . . . , P, where P is the number of selected reference frequencies.

(2) Use the antenna to receive FM and DTMB signals at each sampling point, record the signal strength information on the reference frequency, and construct the RSSI fingerprint database of FM and DTMB signals.

When receiving FM and DTMB signals at the sampling point, due to the large frequency difference between the DTMB signal and the FM signal, an antenna whose receiving frequency range includes these two signals is selected for reception, and multiple measurements are made at each sampling point, and the average value is taken , and calculate its standard deviation.

The measured strength information of different frequencies of FM and DTMB signals is expressed as a vector, and each item of the vector represents the mean value of the signal strength of a frequency. Select N sampling points in the area, and measure the signal strength of P reference frequencies received at each point, then the signal strength of the rth point is expressed as:

为第i个频率进行Q次测量后的平均值，即/>

in

The average value of Q measurements for the i-th frequency, i.e. />

The fingerprint database is jointly constructed from the sampling point position P _r (x _r , y _r ) obtained in step (1) and the signal strength RSS _r and its standard deviation σ _r obtained in this step.

(3) Use the antenna to receive FM and DTMB signals at the point to be located, record the signal strength information on the reference frequency, and obtain the RSSI fingerprint information of the FM and DTMB signals at the point to be located.

In the actual positioning environment, at the point to be located, use the antenna that receives FM and DTMB signals in the receiving frequency range to receive FM and DTMB signals, record the signal strength information on the reference frequency, measure multiple times on each frequency, and take the average value . The RSSI fingerprint information of the point to be measured is expressed as:

为待测点第i个频率进行多次测量后的平均值。in

It is the average value of multiple measurements for the i-th frequency of the point to be measured.

(4) Through the joint positioning algorithm, the RSSI fingerprint information of the FM and DTMB signals of the point to be located is matched with the RSSI fingerprint database to obtain the positioning result.

The joint positioning algorithm combines deterministic information with probabilistic information, and the process of the algorithm is shown in Figure 3.

Deterministic information is obtained by comparing the real measured value with the prior measurement strength of the position. The specific implementation method is: evaluate the Euclidean distance from the point to be located to all the sampling points, and select the nearest K distances. The point corresponding to the distance is the neighbor reference point. At the same time, considering the difference between the K distances and the real-time measured signal strength, different weight values are assigned to the K neighboring reference points. The specific implementation formula is:

为信号强度的先验估计值，LOC_k是近邻参考点的实际位置，K是选择的最近邻的参考点的个数，D_k是待定位点与近邻参考点间的欧氏距离。where rss is a measure of signal strength,

is the prior estimate of the signal strength, LOC _k is the actual location of the neighbor reference point, K is the number of the nearest neighbor reference point selected, and D _k is the Euclidean distance between the point to be located and the neighbor reference point.

In the indoor environment, the signal strength will be affected by Rayleigh fading, but the deterministic information is difficult to resist the influence of Rayleigh fading. Only using this information for positioning is not ideal enough, so the present invention introduces probabilistic information.

Probabilistic information is estimated based on a probability model. The specific implementation method is: when measuring the FM and DTMB signal strength of the sampling point, measure the signal strength of the point multiple times and estimate the signal strength distribution of the point; When , the position of the point to be located is taken as an independent random variable, and the position with the highest probability of occurrence of the measured signal strength is estimated as the positioning result by analyzing the probability of occurrence of the measured signal strength.

According to the Bayesian formula, when the measured signal strength is rss, the probability P(LOC _r |rss) of the positioning position at LOC _r can be calculated as follows:

Among them, P(rss) is the probability that the measured value is rss, P(LOC _r ) is the probability that the point to be located appears in the LOC, and P(rss|LOC _r ) is the probability that the rss intensity appears at the LOC _r point. Considering that the signals at each frequency are independent of each other, the likelihood probability at the sampling point can be expressed by the product of the likelihood probability functions corresponding to k frequencies:

Gaussian distribution is used to fit the distribution of signal strength at each frequency point, that is, P(rss _l |LOC _r ) is expressed as:

Where Q is the actual number of samples measured, and the measured RSS _rq of the location of LOC _r obeys the normal distribution of N(rss _r , σ).

Gaussian regression can be used to estimate the distribution of sampling points affected by Rayleigh fading, and the use of probabilistic information for positioning can overcome the adverse effects of Rayleigh fading, but it cannot make full use of the information between positioning points and neighboring points, so Poor anti-interference ability.

The present invention adopts a joint positioning algorithm and combines deterministic information and probabilistic information. These two types of information have different generation mechanisms, and the errors introduced are also independent of each other. Therefore, the fusion of these two types of information can effectively increase the diversity of positioning information, effectively offset part of the error, and improve the stability of positioning. The fusion formula is:

loc＝ω ₁ loc ₁ +ω ₂ loc ₂ (8)

Among them, loc ₁ is the deterministic information of the positioning result, loc ₂ is the probabilistic information of the positioning result, and ω ₁ and ω ₂ are the weights of the two kinds of information respectively.

The RSSI fingerprint information obtained in step (3) is matched with the RSSI fingerprint database established in step (2), and the deterministic information and probabilistic information of the positioning result are respectively obtained, and then the final positioning result is calculated by using the fusion formula.

The above is only a further description of the present invention, and is not intended to limit the implementation and application of this patent. All equivalent implementations of the present invention should be included in the scope of claims of this patent.

Claims (6)

1. a kind of indoor position fingerprint location method based on FM and DTMB signal, it is characterized in that, by receiving the intensity information of FM and DTMB signal different frequencies, determine the spatial position of indoor environment to be positioned point, comprise the following steps: 1) Select sampling points according to the indoor environment, and select different frequencies of FM and DTMB signals as reference frequencies; 2) Use the antenna to receive FM and DTMB signals at each sampling point, record the signal strength information on the reference frequency, and construct the RSSI fingerprint database of FM and DTMB signals; 3) Utilize the antenna to receive FM and DTMB signals at the point to be located, record the strength information of the signal on the reference frequency, and obtain the RSSI fingerprint information of the FM and DTMB signals at the point to be located; 4) Through the joint positioning algorithm, the RSSI fingerprint information of the FM and DTMB signals of the point to be located is matched with the fingerprint database to obtain the positioning result. 2. the indoor position fingerprint location method based on FM and DTMB signal as claimed in claim 1, is characterized in that, described FM signal is unified by the transmission station of appointment and transmits, and the FM signal that can receive in the same area all comes from same signal launch tower. 3. the indoor position fingerprint location method based on FM and DTMB signal as claimed in claim 1, is characterized in that, described reference frequency is selected based on channel environment. 4. the indoor position fingerprint location method based on FM and DTMB signal as claimed in claim 1, is characterized in that, described FM and the signal strength value of DTMB signal on reference frequency represent with a vector, the rth sampling point This signal strength value of is expressed as

Among them, N is the total number of sampling points selected in the area, P is the number of reference frequencies selected for each sampling point, />

is the mean value of the signal strength after multiple measurements on the i-th reference frequency. 5. the indoor position fingerprint location method based on FM and DTMB signal as claimed in claim 1, is characterized in that, described joint location algorithm has combined the deterministic information of location and probabilistic information, by neighbor reference point, obtain location The deterministic information of the result; the probability information of the positioning result is obtained from the probability distribution of the signal strength; the final positioning result is obtained by assigning different weight values to the two kinds of information. 6. The indoor position fingerprint positioning method based on FM and DTMB signals as claimed in claims 1-5, wherein said method is applicable to any environment where FM and DTMB signals can be received simultaneously.

Images